You are on page 1of 6

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056

Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072

Accurate analysis on hybrid DWT and SVD based Digital Watermarking


for fingerprint security

Zareen Khan, Tirupati Goskula

Mtech student, Electronics and communication, Maharashtra, India


Assistant Professor, Electronics and communication, Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract Identification of persons by way of biometrics There exist two groups of invisible digital image water-
features has evolved significantly over the years. During this marking systems. The first group, spatial watermarking, uses
time, biometric recognition has received much attention due the spatial domain of an original image for watermark
to its need for security. Among the many existing biometrics, insertion, the second group frequency watermarking, uses
fingerprints are considered to be one of the most practical the frequency domain for it. it. The frequency domain based
ones. Techniques such as watermarking been used in attempt watermarking systems are more resistant to a bigger group
to improve security of biometric data. Watermarking is the of possible attacks. In spatial domain, the watermark is
process of embedding information into a carrier file for the embedd into specific pixels of the host fingerprint image. In
protection of ownership/copyright of music, video or image transform domain, the host image is first transformed into to
files. In this paper, to overcome such duplicate things , so a a frequency domain. DWT is the main transform method
new watermarking algorithm is proposed namely Discrete used in transform domain watermarking schemes which also
Wavelet Transform (DWT), Singular Value Decomposition used in JPEG and JPEG2000. Since high frequency
(SVD) The algorithm converted the minutiae into binary components are affected by most of the signal processing
watermark, increasing embedded information capacity. The techniques such as lossy compression, so in order to increase
algorithm can satisfy the transparence and robustness of the the robustness ,the watermark is preferred to be placed in
watermarking system very well and the useful information can the low frequency components. But at the same time human
be extracted accurately even if the fingerprint is severely visual system is very sensitive to changes in low frequency
regarded. range. So in dwt based watermarking technique, the dwt
coefficients are modified to watermark data. A binary image
Key Words: Digital watermarking, Discrete Wavelet or a representation of some data or message can be used as a
Trnasform , Singular Value Decomposition , Arnold watermark. There are procedures in existence that can help
transform, PSNR, CN (Crossing number). to optimize the security of biometric data,onebeing,
information hiding. Information hiding techniques like
1.INTRODUCTION watermarking can add to the security of biometric systems.
Watermarking can be explained as a process of embedding
Biometric systems allow for convenient identification to take information into a carrier file in order to secure copyright
place based on a personsphysical or behavioural typically ownership. Watermarks can be either visible or
characteristics. In comparison with conventional token- non-visible to the human eye. n this study, a wavelet based
based or knowledge based systems, they link identities watermarking algorithm is proposed to enhance the security
directly to the owners. Moreover, these identities cannot be of fingerprint images. The algorithm embeds secret data into
given up or lost easily. The uses of biometric procedures a fingerprint image based on Discrete Wavelet Transform
have evolved rapidly . In todays life, multimedia data (DWT) and Singular Value Decomposition (SVD). The
transmissions over the Internet are growing rapidly. A huge fingerprint image is first converted to the frequency domain
problem with intellectual property laws are solved every and the SVD is applied on both the original fingerprint image
day, because a lot of people download and redistribute the and the watermark image. The singular values (SVs) of the
multimedia data illegally. Plenty of illegal data storages are fingerprint image are then modified with the singular values
spread all over the world. (SVs) of secret image.

Digital watermarking is one of the methods, that can be used


for author's law protection. The basic division of watermarks
is into visible and invisible watermarks. The visible
watermarks are used for example in TV channels, the
invisible watermarks are used to insert information for
example about the author for further proof of ownership to
that work.

2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1208
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072

1.1 FINGERPRINT PATTERNS pattern, no significant up thrusts and the ridges enter on
one side of the impression, and flow out the other with a rise
or wave in the center.

Loops occur in about 60-70% of fingerprint patterns


encountered. One or more of the ridges enters on either side
of the impression, re-curves touches or crosses the line
running from the delta to the core and terminates on or in the
direction of the side where the ridge or ridges entered. Each
loop pattern has is onedelta and one core and has a ridge
count. Radial loops are named after the radius ,bone in the
forearm that joins thehand on the same side as the thumb.

Whorls are seen in about 25-35% of fingerprint patterns


Fig.1.1 Thumb Impression Image representation encountered. In a Whorl, some of the ridges make a turn
through at least one circuit. Any fingerprint pattern which
contains 2 or more deltas will be a whorl pattern. There are
four types of whorl patterns. Plain whorls,central pocket
whorls, Double loop whorls and accidental whorls

1.2 FINGERPRINT MINUTIAE

Fig.1.2 loops Fig.1.3 Whorls

Fig 1.5

A fingerprint is formed from an impression of the pattern


Fig,1.4 Arches
of ridges on a finger. Each person has its own fingerprints
with the permanent uniqueness. So fingerprints have being
A fingerprint consists of three basic patterns of ridges, the
used for identification and forensic investigation for a long
arch, loop and whorl as shown in Figure . An arch can be
time. A fingerprint is composed of many ridges and furrows.
explained as the pattern where ridges begin from one side of
However, shown by intensive research on fingerprint
the finger, ascent in the centre which develops an arc, and
recognition, fingerprints are not distinguished by their
then exits the finger from the opposite side (see Figure a). A
ridges and furrows, but by Minutiae, which are some
loop can be explained as the pattern where ridges begin at
abnormal points on the ridges (Fig.1 (a)). Minutiae points
one side of a finger to create a curve, and are inclined to exit
are most generally the locations of ending or bifurcation of
in the same way they entered (same side- see Figure b). As
ridges in a fingerprint. Every persons fingerprints are
seen above in Figure 10(c), in the whorl pattern, ridges are
unique, and will always maintain their uniqueness
structured in a circular position around a central spot on the
explaining why they have been used for many years for
finger. In general, researchers have discovered that relatives
authentication purposes (Barnes, 2011). Ones fingerprint
frequently share similar fingerprint patterns, which has led to
consists of a pattern of ridges and valleys (located on the top
the concept that fingerprint patterns are genetic.
of the fingertip). The top layer of skin on a finger contains
Arches are actually found in about 5% of fingerprint
the ridges while the lower skin particles contain a pattern of
patterns encountered. The ridges run from one side to the
valleys. The distinctive types of disjunctions in ridges
other of the pattern, making no backward turn. Ordinarily,
(minutiae) hold adequate discriminatory data to distinguish
there is no delta in an arch pattern but where there a delta,
between various fingerprints. Ridge bifurcation (the area
no re-curving ridge must intervene between the core and
where the ridge splits) and ridge ending (the area where the
delta, There are four types of arch patterns: plain arches,
ridge ends) are the most important minutiae points.
radial arches, ulnar arches and tented arches. Plain arches
have an even flow of ridges from one side to the other of the
2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1209
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072

Fig 1.8 Bifurcation

2. WATERMARKING EMBEDDING
Fig 1.6 Minutiae . (Valley is also referred as Furrow,
Termination is also called Ending and Bifurcation is also
2.1 Discrete wavelet tansform (DWT)
called Branch)
Wavelet transform is a time-frequency domain combined
1.3 MINUTIAE POINTS
analysis method. It has multi-resolution analysis features.
Each level of the wavelet decomposition has four sub-images
The major minutia points in a fingerprint consist of: ridge
with same size. Let the LLkstands for the approximation sub
ending, bifurcation, and short ridge as shown in figure.
image andLHk, HLk ,and HHkstand for the horizontal , vertical
and diagonal direction high-frequency detail sub image
respectively. Where the variable k = 1,2,3,...(k N) is the scale
or the level of the wavelet decomposition. DWT is much
preferred because it provides both a simultaneous. spatial
localization and a frequency spread of the watermark within
the host image

The basic idea of discrete wavelet transform in image


process is to multi-differentiated decompose the image into
sub image of different spatial domain and independent
frequencies. After wavelet decomposition, many signal
Fig 1.7 Minutiae points processing, such as compression and filter are likely to
change the high frequency wavelet coefficients. If the
1.2 Ridges- watermark sequence is embedded into this part, its
information may be lost in the processing in sequence, which
A single rolled fingerprint may have as many as 100 or more will reduce the robustness of the watermark [3]. In order to
identification points that can be used for identification ensure the watermark has a better imperceptibility and
purposes. These points of impression can be used as points robustness, the approximation sub-image LL3 coefficients are
of comparison for fingerprint identification. Depending on chosen to embed watermark
how prevalent the ridge characteristics, fewer or more
points of comparison may be needed for positive
identification.

1.3 Bifurcation-

A bifurcation is a point in a finger image from which two


ridges been meet. Bifurcations have the clear appearance of
branch points between curved lines. The number and
locations of the bifurcations and ridge endings, known Fig 2.1 : DWT Decomposition
as minutiae, which are vary from finger to finger in any
particular individual person, and from person to person for 2.2 SINGULAR VALUE DECOMPOSITION (SVD)
any particular finger (for example, the ring finger on the
right hand). When a set of fingerprint images is obtained If a m*n image is represented as a real matrix A , it can be
from an individual, the number of minutiae is recorded for decomposed as:
each finger. A bifurcation is a point in a finger image from
which two ridges been meet. A= U S VT

It is called a singular value decomposition of A . Where U is a


m*m unitary matrix, S is a m*n matrix with nonnegative

2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1210
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072

numbers on the diagonal and zeros on the off diagonal, and


VT denotes the conjugate transpose of V, an n*n unitary START
matrix. The nonnegative components of S represent the
luminance pixel value of the image. Changing them slightly
does not affect the image quality and they also dont change Read host Fingerprint Image
much after attacks, watermarking algorithms make use of
these two properties.
Fingerprint Pre- Implement 3-
2.3 ARNOLD TRANSFORM
processing level wavelet
Arnold transform is commonly known as cat face transform, transform
it randomizes the original organization of pixels in an (enhancement,
original image. However, if iterated enough times, the binarization, Select
original image reappears. Arnold transform is the position
shift of one to-one point. Arnold transformation on digital
Thinning) approximation
image of size NN defined by Eq. 1 is a one-to-one LL3 Sub-band
transformation Minutiae
Image(A)
extraction

Conversion of Implement SVD


.. (1) minutiae points on A i.e; A= USVT
Where i, j {0, 1, . . . .,N-1}. Where (i, j) is the location to binary
coordinates of the original image pixels and (i, j) is the watermark (W)
location co-ordinates of image pixels that after transform
When all the coordinates are transformed, the image we
obtain is scrambled images. Arnold
Transform
3. PROPOSED WATERMARKING ALGORITHM
Obtain watermark image coeffiecient Aw as S + W
= UW SW VWT
3-level 2D DWT SVD
decomposition AW = USW VT

Compare Apply reverse wavelet transform


Arnold Svs of
Transform watermark
Fingerprint Change double precision real number to unsigned
with AW = USW
Pre- 8-bit integer
original
Processing
image
Watermarked fingerprint
IDWT
Conversion
Minutiae Fig..3.2 flow diagram of Watermarking (Minutiae)
of
Extraction(x,y) Embedding.
minutiae Watermarked
to binary fingerprint The proposed watermarking scheme has been shown in
watermark Fig. 3.The cover image is the fingerprint while the
watermark is a binary image equivalent to the minutiae of
the cover fingerprint. The scheme has been divided into
Fig.3.1 Block Diagram of Watermarking Embedding two sections:
System
A. Watermark (Minutiae) Embedding

2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1211
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072

B. Watermark (Minutiae) Extraction Step 4- Perform Arnold transform for watermark image W.
which is shown in the following figure-
3.1 Watermark (Minutiae) Embedding scheme

The steps of embedding watermarks can be described as


follows:
Step1-The fingerprint image is decomposed into its 3-level
two-dimensional DWT coefficients. Out of the all sub-
bands, only LL3 approximation sub-band is selected (denotes
as A ).
Step 2- Fingerprint Preprocessing
A real fingerprint might have discontinuities that might lead Fig. 3.3 Arnold transform watermark
to spurious minutiae. Therefore, minutiae extraction is
preceded by fingerprint preprocessing, which involves Step 5- We obtain the watermarked image coefficients
normalization, ridge orientation and frequency estimation. matrix AW through the following three steps:
Finally, the ridge orientation and frequency estimation
values are used for filtering the fingerprint using Gabor 1. A = USVT
wavelet. Gabor filtering enhances the ridges oriented in the 2. S + W = UW SW VWT
direction of the local orientation, and decreases anything 3. AW = USW VT
oriented differently. Hence, the filter increases the contrast
between the foreground ridges and the background, whilst Step 6-Apply reverse wavelet transform for original image,
effectively reducing noise. The filtered output is then and then changing the double-precision real number to
binarized and thinned to one-pixel width. unsigned 8-bit integer. Thus, obtain the watermarked image
Step 3-Minutiae Extraction in which watermark are embedded.
Minutiae points such as end points and bifurcation points are
identified by calculating Crossing number (CN). CNvalue 3.2 Watermark (Minutiae) Extraction scheme
isdefined as half the sum of the differences between pairs of
adjacent pixels in the eight neighborhoods.
CN=0.5Pi_Pi+1 |, P1= P9 START
where Pi is the pixel value in the neighborhood of a pixel P.
If crossing Number is 1, 2 and 3 or greater than 3 then
minutiae points are classified as Termination, Normal ridge
Watermarked Image
and Bifurcation respectively. it is shown as
3-level wavelet transform

Select LL3 Sub-band

Fig.a Crossing number=2 Normal ridge pixel


Apply SVD on A*, A*=U*S *V*T

Obtain D*=UWS*VWT

Fig.b Crossing number=1 Termination Obtain embedded watermark as W= (D-S)

Change double precision real number to


unsigned 8-bit integer

Fig. 3.4 Flow diagram of Watermark (Minutiae) Extraction


Fig.c Crossing number=3 Bifurcation points

2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1212
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072

We can extract the watermark by the reverse [4] Prabhishek Singh, R S Chadha, "A Survey of Digital
Calculation of watermark extraction: Watermarking Techniques, Applications and Attacks",
International Journal of Engineering and Innovative
Step 1-Perform a 3-level wavelet transform using haar Technology (IJEIT), Volume 2, Issue 9, ISSN: 2277-3754,
wavelet for watermarked image, and obtain low-frequency March 2013.
wavelet coefficient LL3 (denotes as A*). [5] Robert, L., and T. Shanmugapriya. "A study on digital
Step 2-Apply SVD to the A*, such that A* =U * S1 *VT*, and watermarking techniques." International Journal of
obtain U *,S1* and VT*. Recent Trends in Engineering , Vol. 1, no. 2 , pp.223-225,
Step 3: Now by using values of U w, VW and S1 *, obtain D* 2009.
according D*= UW S1 *V T, in the end we canobtain the [6] Reddy, R., M. V. N. Prasad, and D. S. Rao. "Robust Digital
watermark which is embedded according to Watermarking of Images using Wavelets." International
W* = (D* S) /. Journal of Computer and Electrical Engineering, vol.1,
Step 4-Finally changing the double-precision real number to no.2, pp.1793-8163, 2009.
unsigned 8-bit integer for watermark image, and perform [7] Mohammad Abdullatif, AkramM.Zeki, JalelChebil, Teddy
inverse Arnold transform for watermark image. Surya Gunawan, Properties of Digital Image
Step 5- Minutiae points are then regenerated by stacking bit Watermarking,IEEE 9th International Colloqium on Si.
planes and converting them back to decimal system. [8] PrafulSaxena, ShanonGarg and ArpitaSrivastava, DWT-
SVD SemiBlind Image Watermarking Using High
4. CONCLUSION Frequency Band," 2nd International Conference on
Computer Science and Information Technology ,
In this paper, we proposed a watermarking algorithm for Singapore, April 28-29, 2012.
fingerprint images based on hybrid DWT and SVD domain. [9] I. Cox, J. Kilian, F. T. Leighton, and T. Shamoon, (1997).
The purpose of the proposed watermarking scheme is to Secure spread spectrum watermarking for multimedia,
extract minutiae from even tampered fingerprint. Applying IEEE Transactions on Image Processing, Vol. 6, no. 12,
the Arnold transform to the watermark makes the results pp. 1673-1687.
even better. [10] Akshay Kumar Gupta, and Mehul S Raval, "A robust and
secure watermarking scheme based on Singular values
ACKNOWLEDGEMENT replacement,"in Indian Academy of Sciences, vol. 37,
Part 4, August 2012, pp.425-440.
Its an honor for me for having such a Great Guide. I would
like to thank my guide Prof. Tirupati Goskula for all of their
help. They introduced me to the topic of Scientific Analysis
on hybrid DWT and SVD based Digital Watermarking for
fingerprint security and were always supportive. I am
very thankful to my Guide Prof. Tirupati Goskula,
Assistant Professor in Electronics and Communication
Engineering for his valuable support, suggestions and advice.
I would not have been able to write paper without their
assistance.

REFERENCES

[1] AsmaFurqan ,Munish Kumar ,"Study and Analysis of


Robust DWT-SVD based Digital Image Watermarking
technique using MATLAB ", 2015 IEEE International
Conference on Computational Intelligence and
Communication Technology.
[2] Onur Jane ,HakkiGokhanIk , ErsinElbasi, "A Secure and
Robust Watermarking Algorithm Based on the
Combination of DWT,SVD and LU Decomposition with
Arnold's Cat Map Approach ", 2013 IEEE International
Conference on Electrical and Electronics Engineering
(ELECO).
[3] ShuchiSirmour, ArchanaTiwari , "Hybrid DWT-SVD based
Digital Watermarking Algorithm for Copyright
Protection" , March 2014 International Journal of P2P
Network Trends and Technology (IJPTT).

2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1213